use of different reference foods in determining the glycemic index of starchy and non-starchy test foods

BACKGROUND: Glycemic index (GI) is intended to be a property of food but some reports are suggestive that GI is influenced by participant characteristics when glucose is used as a reference. OBJECTIVE: To examine the influence of different reference foods on observed GI. DESIGN: The GIs of five varieties of rice and a sugary beverage (LoGiCane™) were tested in 31 European and 32 Chinese participants using glucose or jasmine rice as reference foods. The GIs of two ready-to-eat breakfast cereals (Kellogg’s cornflakes and Sustain) were tested in 20 younger and 60 older people using glucose or Sustain as reference foods. RESULTS: The GIs of rice tended to be higher in the Chinese compared with the Europeans when glucose was used as a reference (jasmine 80 vs 68, P = 0.033; basmati 67 vs 57, P = 0.170; brown 78 vs 65, P = 0.054; Doongara 67 vs 55, P = 0.045; parboiled 72 vs 57, P = 0.011). There were no between-group differences in GI when jasmine rice was the reference. The GIs of breakfast cereals tended to be lower in younger compared with older groups (cornflakes 64 vs 81, P = 0.008; Sustain 56 vs 66, P = 0.054). There was no between-group difference in the GI of cornflakes when Sustain was the reference (cornflakes 115 vs 120, P = 0.64). There was no ethnic difference in GI when glucose was the reference for another sugary food (LoGiCane™ 60 vs 62; P = 0.69). CONCLUSIONS: A starchy reference may be more appropriate than a glucose beverage when attempting to derive universally applicable GI values of starchy foods. TRIAL REGISTRATION: The Chinese/European trial is registered with the Australian New Zealand Clinical Trials Registry as ACTRN12612000519853 .